Atomistic models are a particularly valuable component of contemporary investigations of the mineral-water interface because they provide mechanistic insights that are difficult to discern from laboratory-based observations alone. While the timescales accessible by simulation are often sufficient to characterize relatively small amplitude fluctuations in the structure of a solvent, electrolyte solution, or mineral-water interface, more infrequent larger scale fluctuations that may lead to reactive events such as the detachment / attachment of ions at mineral surfaces are often inaccessible. Yet, it is these “rare events” that we are most often interested in, because these processes are ultimately responsible for the chemical behavior we observe at the macroscale. Among the many methods that have been introduced to simulate rare events, the reactive-flux (RF) and transition path sampling (TPS) approaches standout as the most widely utilized for computing reaction rates. Although TPS is not burdened by the necessity of defining a good reaction coordinate (such as is needed for RF-based approaches) the computational overhead associated with the method is significant, and accounting for barrier recrossing events leads to slow convergence of the calculated rate constants. One recently evolved alternative to standard TPS simulations is the Forward Flux Sampling (FFS) approach, which improves the computational efficiency of path sampling simulations by reformulating the rate constant in terms of the effective positive flux. Unfortunately, this increased efficiency comes at the price of reintroducing the requirement of a prescribed reaction coordinate and all its associated shortcomings. The main objective of this work is to eliminate the need to define a reaction coordinate in FFS simulations, thus taking advantage of FFS’s improved computational efficiency, and enabling the obtention of unbiased dynamical trajectories a la TPS. To meet this aim I hypothesize that the FFS interfaces can be redefined in terms of elapsed time rather than a geometrically defined measure of reaction progress. To test this assertion, I have applied this method to investigate ligand exchange reactions (in solution and at mineral surfaces) whose kinetics and mechanisms are well-constrained by experimental observations. Herein I contrast these simulation results with experiment and other rare events sampling techniques.     <br/>

Development of rare events simulation tools to probe mineral surface reactivity

253rd National Meeting (2017)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.